Best AWS Services and Cost Estimation Strategy for a Multi-Tenant Application with Individual Databases

Question

Hello AWS Community,

I am in the process of planning a unique application deployment that involves multiple clients, each with their personalized version of the same application. For every client, we plan to launch a new instance of the application under a specific subdomain and pair it with a dedicated database to ensure isolation and facilitate management. I'm seeking guidance on the best AWS services and strategies to deploy this scenario efficiently and cost-effectively.

Here are the key details to consider:

We expect to support around 120,000 active users.
Anticipated peak usage is up to 10,000 concurrent users.
The application will handle approximately 150 read requests per second and 50 write requests per second.
Requests will be relatively small but will all interact with the database.
Given these requirements, I'm looking for advice on:

The most suitable AWS services for hosting the application instances and databases, considering the need for scalability and isolation between clients.
Strategies for cost estimation and optimization, especially in terms of database management and handling variable traffic with efficiency.
Any insights, experiences, or recommendations on how to structure this deployment and estimate costs would be greatly appreciated. Thank you for your help!

Gary Mclean · Answer

There isnt enough information in the post to guide and design your architecture. A Solutions Architect would need a deeper dive into your application and how it operates to help design and answer the questions. I can think of handful of questions.

AWS Has over 200 services and growing constantly and depending on your requirements would dictate the solution.

You may wish to engage with a SA either directly with AWS or a partner such as where I am based

Antonio_Lagrotteria · Answer

I agree that more context is needed in regards of traffic patterns and more, but given the numbers you gave, you probably better off to start with Serverless components such as lambda functions, API gateways and learn from there.

From database perspective, depending on access patterns, you may look into either dynamo or RDS. At this stage Aurora feels a bit of overhead, though you haven’t mentioned anything about how global your app is.

If it is of course you may want to look into Cloudfront and s3 for hosting front ends.

Above components are fairly cheap for traffic you mention and can easily be managed via aws budgets, and cost explorer.

Best AWS Services and Cost Estimation Strategy for a Multi-Tenant Application with Individual Databases

Relevant content